2009-08-012024-05-18https://scholars.lib.ntu.edu.tw/handle/123456789/707076摘要：電子可靠度泛指與電子產品有關之可靠度，包括電路、材料、封裝、散熱、降噪、減振、減少應力集中、耐久性、壽命等廣泛議題，目前機械人員較常參與的研究偏重於封裝可靠度，所採用的方法幾乎著重於力學分析與材料性質探討，較少引用可靠度工程所系統化介紹的分析方法或如MIL-HDBK-217f、Telcordia SR-332等電子可靠度規範與資料庫探討電子可靠度問題，有鑑於此，本計畫擬結合封裝力學與可靠度工程相關分析方法，探討電子封裝的可靠度。而基於覆晶晶片尺寸封裝(Flip-Chip Chip Scale Packages, FCCSP)技術在記憶卡、IC卡、手機SIM卡、通訊晶片或一般門禁非接觸性的感應卡等熱門電子產品上之廣泛應用，本計畫擬進行一為期兩年的研究，應用數值模擬方法，配合Garofalo-Arrhenius潛變方程式及修正後的Coffin-Manson疲勞壽命預估公式，預測一採無鉛錫球焊接之FCCSP於熱循環負載下，FCCSP無鉛錫球接點之可靠度(Solder Joint Reliability, SJR)。我們首先將以ANSYS 11.0有限元素軟體為工具，分析FCCSP於熱循環負載下之力學行為；而後納入全域及局部模型(Global Model/Submodel)，並比較兩種分析結果之差異及其適切性；接著，搭配ANSYS之參數化設計語言(ANSYS Parametric Design Language, APDL)與機率設計系統模組(Probabilistic Design System, PDS)，改變FCCSP結構尺寸和材料參數，探討尺寸和材料不確定性對FCCSP疲勞壽命分佈與可靠度之影響，也評估修正後Coffin-Manson疲勞壽命預估公式中，係數不確定性對FCCSP疲勞壽命分佈與可靠度之影響。我們也將依MIL-HDBK-217f或Telcordia SR-332等規範方法及其資料庫，計算FCCSP在特定環境下之失效率，並與前述有限元素分析的結果相互比對、驗證。最後，我們將應用可靠度工程理論及ANSYS設計最佳化模組(Design Optimization)，進行FCCSP與無鉛錫球相關參數的最佳化機率設計。<br> Abstract: Researches related to electronic reliability may be divided into two major categories. The first one follows the guidance of MIL-HDBK-217 and applies statistical methods to find failure probabilities or failure rates of electronic devices or components. The other one employs engineering techniques to study failures of devices or components based on physical laws and tries to modify product designs for longer lives. The recent popular electronic package reliability study belongs to the second one. It is felt that the gap between two categories should be narrowed. Therefore, in the present study, we propose to study, from both mechanics and statistical points of view, the reliability of electronic packages. The type of Flip-Chip Chip Scale Packages (FCCSP) is selected in our demonstrative analysis since it is widely used in modern electronic devices. First, we will apply finite element method (FEM) to find stress, strain and/or dispassion energy of a certain type of FCCSP. The life of the package can be evaluated through a modified Coffin-Manson law or its equivalents in consideration of the temperature cycling environment. In the above analyses, if there is uncertainty in any of the involved parameters, the life of the package becomes a random one, and reliability techniques can be employed to find the quantitative reliability of the package. The integrated analysis stated above, in fact, starts from consideration of the so-called physics of failure (PoF). Nevertheless, it can be linked directly or indirectly to failure rates as used in MIL-Hdbk-217 as well as many other reliability codes such as Telcordia SR-332, which are resulted primarily from statistical analysis, since quantitative reliability analysis has been adopted in the mechanics analysis. Aside from the above demonstrative analysis, FEM modules such as ANSYS Parametric Design Language (APDL) and Probabilistic Design System (PDS) will be examined and employed in the proposed study to see if they are helpful in the analyses. The final purpose of the study is, of course, to find an optimal and robust design process for FCCSP. The process also coincides with the so-called probabilistic engineering design.電子可靠度失效率覆晶晶片尺寸封裝全域/局部模型分析方法機率設計系統模組。Electronic ReliabilityFailure RateFlip-Chip Chip Scale Packages (FCCSP)Global Model/Submodel Finite Element MethodProbabilistic Design System (PDS).